Which term refers to machine learning systems that use explicit, human-understandable rules for reasoning?

The term that describes machine learning systems utilizing explicit, human-understandable rules for reasoning is symbolic systems. These systems rely on a formal set of rules and symbols to represent knowledge and make decisions, allowing for a clear traceability of the reasoning process. This approach is particularly beneficial in contexts where transparency and interpretability are crucial, as it allows humans to understand how conclusions are reached based on the rules encoded in the system.

In contrast, statistical models focus on identifying patterns in data through probabilistic methods and often lack the explicit interpretability found in symbolic systems. Neural networks operate on complex architectures that learn from data through nonlinear transformations, making them powerful but often opaque in terms of reasoning. Generative models aim to create new data samples from learned representations, but do not inherently focus on human-understandable reasoning rules. Thus, symbolic systems stand out as the choice that emphasizes human-readable reasoning in contrast to other machine learning methodologies.